Copyright Warning & Restrictions
The copyright law of the United States (Title 17, United
States Code) governs the making of photocopies or other
reproductions of copyrighted material.
Under certain conditions specified in the law, libraries and
archives are authorized to furnish a photocopy or other
reproduction. One of these specified conditions is that the
photocopy or reproduction is not to be “used for any
purpose other than private study, scholarship, or research.”
If a, user makes a request for, or later uses, a photocopy or
reproduction for purposes in excess of “fair use” that user
may be liable for copyright infringement,
This institution reserves the right to refuse to accept a
copying order if, in its judgment, fulfillment of the order
would involve violation of copyright law.
Please Note: The author retains the copyright while the
New Jersey Institute of Technology reserves the right to
distribute this thesis or dissertation
Printing note: If you do not wish to print this page, then select
“Pages from: first page # to: last page #” on the print dialog screen
The Van Houten library has removed some of the
personal information and all signatures from the
approval page and biographical sketches of theses
and dissertations in order to protect the identity of
NJIT graduates and faculty.
ABSTRACT
BIAXIALLY LOADED SLENDER REINFORCED CONCRETE
COLUMNS SUBJECTED TO SUSTAINED LOADS
by
Song Gu
A generalized analytical approach is presented in this research to predict the behavior
both of slender and short reinforced concrete columns under sustained biaxial
eccentric load.
The present analysis proposes equations established at a cross section of a
reinforced concrete column by combining force equilibrium, constitutive law, and
compatibility conditions. The strain and curvature of each section and the deflection
of the column can then be obtained and resolved.
The established creep computation models, recommended separately by
American Concrete Institute (ACI) 209R92 and the Comite' EuroInternational du
Be'ton (CEB)FIP 1990 Model Code have been used to calculate creep and shrinkage
for a member under a constant elastic compressive concrete strain for a given period.
This present analysis also proposes a computerized method for time and strain
adjustment. The Time and Strain Adjustment of Creep Method, combining a creep
calculation with a constant elastic strain such as those mentioned above, the creep
strain at each cross section can then be calculated, stored and adjusted to age of
concrete, load changes and deflection modifications during each time increment
phase.
In the conventional loaddeflection analysis process, with projected
transformations, a spatial deflection curve is resolved into a couple of planar curves
located separately in two orthogonal plans. Based on the force equilibrium equations
of inner force at a column section, a set of three simultaneous nonlinear differential
equations are derived to establish the relationships between the planar curve functions
with the eccentric load upon the top of column. Using the Green's Integral Formula,
the strain and stress nonlinear functions and column section properties can be solely
integrated into a few important coefficients of the differential equations. Thus, it
makes the approach also suitable for columns with nonrectangular sections and any
kinds of constitutive laws of materials.
The presented rational computer analysis results have been compared with the
existing biaxial and uniaxial experimental data, which are available in literature.
They indicate that the results from the proposed analysis correlate with experimental
data well.
BIAXIALLY LOADED SLENDER REINFORCED CONCRETE
COLUMNS SUBJECTED TO SUSTAINED LOADS
by
Song Gu
A Dissertation
Submitted to the Faculty of
New Jersey Institute of Technology
In Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy in Civil Engineering
Department of Civil and Environmental Engineering
May 2002
Copyright © 2002 by Song Gu
ALL RIGHTS RESERVED
APPROVAL PAGE
BIAXIALLY LOADED SLENDER REINFORCED CONCRETE
COLUMNS SUBJECTED TO SUSTAINED LOADS
Song Gu
Dr.
C. T. Thomas Hsu, Dissertation Advisor Date
Professor of Civil and Environmental Engineering, NJIT
Dr. Methi Wecharatana, Committee Member Date
Professor of Civil and Environmental Engineering, NJIT
Dr. Dorairaja Raghu, Committee Member Date
Professor of Civil and Environmental Engineering, NJIT
Professor Walter Konon, Committee Member
Date
Professor and Associate Chair of Department of Civil and Environmental
Engineering, NJIT
Dr. Michael Y. Xing, Committee Member Date
Associate of ThortonTomasetti Engineers, Newark, NJ
BIOGRAPHICAL SKETCH
Author: Song Gu
Degree: Doctor of Philosophy
Date: May 2002
Undergraduate and Graduate Education:
• Doctor of Philosophy in Civil Engineering
New Jersey Institute of Technology, Newark, NJ, May 2002
• Master's Degree in Structural Engineering
Tongji University, Shanghai, P. R. of China, June 1988
• Bachelor's Degree in Structural Engineering
Tsinghua University, Beijing, P. R. of China, July 1985
Major: Civil Engineering
Publications:
Song Gu and C. T. Thomas Hsu (2002), "Computer Analysis of Reinforced Concrete
Columns Subjected to Biaxial Sustained Loads" 15
th ASCE Engineering
Mechanics Conference, Columbia University, NY, June.
Song Gu, Bolong Zhu and Yuzhou Chen (1989), "Study on Aseismic Detailing of the
Reinforced Concrete Frame Connections", Journal of Structural Engineers, No. 2,
pp. 1824, Shanghai, China (in Chinese).
iv
This dissertation is dedicated to my whole family members both in United States
and the People's Republic of China
ACKNOWLEDGEMENT
I wish to express my sincere gratitude to my Advisor, Professor C. T. Thomas Hsu, for
his guidance, encouragement, and help throughout the course of this research.
Special thanks to my dissertation committee members, Professor Methi
Wecharatana, Professor Dorairaja Raghu, Professor Walter Konon, and Dr. Michael Xing
for their constructive evaluation and valuable suggestions. Thanks also to Professor
Edward G. Dauenheimer, for his evaluation and suggestions in this study.
Appreciation is expressed to Mr. Russell Sage, PE and the company of NK
Architects for their understanding and encouragement.
I wish to thank my friend, Dr. Xianglin Gu, for his suggestions and
encouragement.
Also to thank my parents, Boqi Gu and Ziwen Qian, for their hopes,
encouragement and assistance.
Finally, I wish to express my love to my wife Xun Chen and my children for their
constant encouragement and understanding, throughout the course of this Ph.D. study.
vi
TABLE OF CONTENTS
Chapter Page
1 INTRODUCTION..
1
1.1 Research Background
1
1.2 Objectives of Research
3
1.3 Statement of Originality
4
1.4 Outline of Research
5
2 LITERATURE REVIEW 7
2.1 Classification of Concrete time dependent properties
7
2.2 Mechanism of concrete Creep and Shrinkage....
8
2.2.1 Factors Affecting Creep and Shrinkage
8
2.2.2 Mechanism of Concrete Creep
12
2.2.3 Mechanism of Concrete Shrinkage
14
2.2.4 Different Properties of Creep and Shrinkage between NSC and HSC... 15
2.3 Analytical Models of Concrete Creep and Shrinkage
16
2.3.1 ACT 209R92 Models
16
2.3.2 Models of CEBFIP Mode Code 1990
19
2.3.3 Nonlinear Model for Creep
22
2.3.4 Microstructure Model for Creep 23
2.4 Studies on Slender RC Column Buckling under Sustained Loads
24
2.4.1 Effects of Creep and Shrinkage on Strength of RC Column
24
2.4.2 Experimental Studies
25
vii
TABLE OF CONTENTS
(Continued)
Chapter Page
2.4.3 Theoretical Analysis
26
2.5 Engineering Approach Specified by Engineering Design Codes
32
2.6 Summary 35
3 PROPOSED ANALYTICAL APPROACH
37
3.1 Analysis Strategies
37
3.2 Analysis Assumptions 39
3.3 Solution Procedures
39
3.3.1 Basic Equations 39
3.3.2 Iteration Controlled Derivative Equations
43
3.3.3 Time and Strain Adjustment of Creep Method
48
3.3.4 Constant Step of Creep Increasing Technique
50
3.3.5 Main Algorithm Flow Chart
51
3.4 Convergence and Stability of Solution
51
3.5 Limitations of Proposed Analytical Model
53
4 TEST DATA VERIFICATION
55
4.1 Slender Columns under Sustained Uniaxial Bending Loads
55
4.1.1 Test Data from Goyal et al. (NSC)
55
4.1.2 Test Data from Drysdale et al. (NSC)
59
4.1.3 Test Data from Claeson et al. (HSC)
61
4.2 Slender Columns under Sustained Biaxial Bending Loads
64
4.2.1 Test specimens
64
viii
TABLE OF CONTENTS
(Continued)
Chapter
Page
4.2.2 Test Procedure and Equipment
66
4.2.3 Comparison of Test Results with Numerical Analysis
69
4.3 Discussion of Selection of Ultimate Creep Coefficient Value
73
5 DESIGN RECOMMENDATIONS 75
5.1 Design Examples: Comparison with Current ACI Design Codes 75
5.1.1 Ultimate Strength of Slender Concrete Column under Uniaxial
Bending Load
75
5.1.2 Ultimate Strength of Slender Concrete Column under Biaxial
Bending Load
80
5.2 Mechanism of the RC Column Strength
82
5.3 Estimation of Lateral Deflection of Biaxially Loaded RC Slender Column .. 88
5.4 Recommendations.
95
5.4.1 Engineering Practice
95
5.4.2 Future Research
96
6 CONCLUSIONS
97
APPENDIX A FACTORS AND COEFFICIENTS OF
ACI 209R92
MODELS (1994)
100
APPENDIX B COEFFICIENTS OF
CEBFIP MODEL CODE 1990 (1993)
103
ix
TABLE OF CONTENTS
(Continued)
Chapter
Page
APPENDIX C ELASTIC ANALYSIS VERIFICATION RESULTS
106
APPENDIX D PROCESS OF DEFLECTION SOLUTION
108
APPENDIX E NOTATIONS OF SOURCE CODE AND FUNCTIONS OF
PROGRAM C
112
E.1 Notations of Source Code
112
E.2 Functions of the Program 117
E.3 Detailed Algorithm Flow Chart
118
REFERENCES
124
LIST OF FIGURES
Figure
Page
2.1 Changes in Strain of A Loaded and Drying Specimen
7
2.2 Relationships between Creep and Various Factors Which Influence Creep 11
2.3 Typical Relationships between Deformations of Elastic, Creep and the
Recoveries 17
2.4 Standard Shrinkage Strain Variation with Time after Moist Curing 19
2.5 A Nonlinear Model for Concrete Consisting of Two Springs and One Dashpot 22
2.6 Modified Superposition Method for Calculating Concrete Creep 31
3.1 Projected Curvatures of Ø, and Φy
38
3.2 System of Global Forces and Deflections
39
3.3 Concrete Compressive Stressstrain Relationship 42
3.4 Main Flow Chart of the Algorithm 51
4.1 Reinforcement Details of Specimens of Goyal et al. 55
4.2 Variation of y with Time at Different Stress Levels 57
4.3a Deflectiontime Curve of Column H
58
4.3b Deflectiontime Curve of Column G 58
4.3c Deflectiontime Curve of Column R
59
4.4a Deflection Results of D.2.A
60
4.4b Strain Diagrams of D.2.A
61
4.5 Details of Columns and Test Equipment in Claeson et al.'s Experiment
62
4.6 Details of Reinforcement in Claeson et al's Experiment
62
xi
LIST OF FIGURES
(Continued)
Figure
Page
4.7 Deflection Results of H201 64
4.8 Column Details of Specimens in Drysdale et al.'s Experiment 66
4.9
Creep Test Results from Drysdale et al.'s Experiment 67
4.10a Deflection Results of B 2 A 68
4.10b Strain Diagrams of B.2.A 69
4.11 Strain Diagrams of B.3.B 71
4.12 Concrete Maximum Elastic Strain History of B.2.A
72
4.13 Concrete Maximum Elastic Strain History of B.3.B 72
5.1 Details of Columns in Design Examples
76
5.2 Short Column PM Interaction Diagram
76
5.3 Effects of Slenderness 77
5.4
Effects of Eccentricities. 78
5.5 Effects of Longitudinal Reinforcement Ratios 79
5.6
Effects of Concrete Compressive Strength
80
5.7
Effects of Eccentricity Orientations to Columns with Square Cross Sections
81
58 Effects of Eccentricity Orientations to Columns with Rectangular Cross
Sections 82
5.9 Slender Column PM Interaction Diagram 84
5.10 Primary and Secondary Moments and Deflections
85
XI'
LIST OF FIGURES
(Continued)
Figure Page
511 Comparison of equations for EI with EI values from MomentCurvature
Diagrams for ShortDuration Loading
93
C.1 Column Details 107
E.1a Detailed Algorithm Flow Chart I 119
E. 1 b Detailed Algorithm Flow Chart II 120
E. 1 c Detailed Algorithm Flow Chart III 121
E.1d Detailed Algorithm Flow Chart IV
122
E. 1 e Detailed Algorithm Flow Chart V 123
LIST OF TABLES
Table
Page
4.1
Aggregate Grading of Goyal et al.'s Experiment
56
4.2 Column Details and Ultimate Loads
56
4.3
Composite of Concrete Mixtures, kg/m
3 63
4.4
Hardened Concrete Properties at Different ages for Each Concrete Strength.. ...
63
C.1 Column Specimens
106
C.2
Maximum Moment Results (kipsin)
106
C.3
Maximum Load Deflection Results
107
CHAPTER 1
INTRODUCTION
1.1 Research Background
The study of the behavior of slender reinforced concrete columns under sustained loads
has experienced a history of nearly 50 years.
Ostlund (1957) from Norway is probably the earliest researcher in the world to
study the strength capacity of slender reinforced concrete columns under sustained loads.
He suggested using the reduced elastic modulus of concrete to estimate creep effects.
This concept established the basic direction for the engineering approach afterwards.
From the 1960s, many scholars in North America appeared to be interested in this
research area. Mauch & Holley (1963) at M.I.T., performed theoretical analysis on creep
buckling of reinforced concrete columns and their work became the first publication
appeared in North American academic journals. Just at the same time, ACI Building
Code (ACI Committee 318, 1963) required engineering consideration of creep effects on
deflections of slender columns under sustained loads in some situations. This appeared
the first reinforced concrete (RC) structures building code to specify such effects in
engineering design. Eight years later, in the next generation version of ACI Building
Code (ACI Committee 318, 1971), a formal design method for slender column under
sustained loading was introduced.
Drysdale and Huggins (1971) at University of Toronto, Toronto, Canada,
published their study on sustained biaxial load on slender concrete columns. This is the
first publication in this field to study the slender concrete columns subjected to biaxial
2
bending moments and axial loads. Their experimental data of biaxially loaded columns
under sustained loads are still the only test results available in the literature at this time.
Recent advance in high strength concrete (HSC) structures has resulted in using
smaller section sizes of columns. As the time dependent properties of HSC are different
from those of normal strength concrete (NSC), Claeson & Gylltoft (2000) carried out their
experiment on slender HSC column under sustained loads. In their analysis, the method
provided by CEBFIP Model Code was used.
According to literature research, the most active research period in this field
happened during 19631971. During that period, there were more than 10 research
papers being published in academic journals in North America and a formal design
method, named as
Moment Magnifier Method (MMM), was introduced into the ACI
building code (1971). Today, this MMM is still being used in the latest ACI building
code (1999).
The argument to MMM is that, the quantity of concrete timedependent properties
has not being considered, which means, the capacity of slender column strength is
unchanged no matter how big creep and shrinkage are being developed. This is probably
due to the consideration of simplicity and feasibility in using formulas. On the other
hand, it has reflected the fact that, the concrete timedependent properties are still not
fully understood at this time. In addition, due to time factor, this type of analysis becomes
fourdimensional analysis assuming that the columns are under biaxial bending loads.
For creep, the time effects are definitely not a linear. Thus, it was very difficult to
perform such an analysis with a typical Personal Computer or even a Work Station in ten
or twenty years ago, which would cost much time and require much storage.
3
Since 1971, more and more research results have been published on time
dependent properties of concrete. Many theories and analytical methods have been
established and updated to describe the behavior of concrete creep and shrinkage. ACI
Committee 209 and CEBFIP have developed theoretical analysis models for concrete
creep and shrinkage (ACI Committee 209, 1994, & CEBFIP, 1993). Although these
theories of concrete timedependent properties are not perfect, they do provide powerful
theoretical assistance to the study of behavior of long RC columns under sustained loads.
In recent years, the highspeed development of computer technology also makes
it possible to overcome the storage and speed difficulties when tackling three or four
dimensional nonlinear structural problems.
In engineering practice, it has passed more than a century since the concrete
material was invented. Thousands of skyscrapers in concrete structures have been built
up all over the world. Among those concrete structures, a lot of them have been standing
for more than 50 years. Some of them have stood for even more than 80 years, which are
close to the end of concrete life duration. All these facts have raised a question that needs
to be solved which is, what is really happening to the behavior of those slender columns
inside those old buildings.
1.2 Objectives of Research
The objectives of this research are:
1. To establish a theoretical analysis model and algorithm in order to simulate the
loading and deformation process of slender columns made of NSC/HSC under bi
axial sustained loads (uniaxial bending is treated as a special case of biaxial
4
bending). Existing experimental data are to be used to verify the validity of analytical
models developed.
2.To analyze important material and geometry variables that influence the behavior of
slender columns under sustained loads with the established method.
3.To compare and evaluate the current design criteria with design examples.
1.3 Statement of Originality
A generalized theoretical approach is created to predict the behavior of slender reinforced
concrete columns made of NSC/HSC under sustained biaxial/uniaxial eccentric loads.
This approach includes the following original points:
1.
Applying Green's Integral Transformation to the three balanced force equations, the
strain and stress relationship and column section properties are integrated into a few
important coefficients of the differential equations. Thus, it makes the algorithm
generalized to nonrectangular section columns and all kinds of constitutive
functions.
2.Established a computerized method for the adjustment of creep increments due to the
changes of elastic strain during the sustained loading duration. The method is named
as Time and Strain Adjustment of Creep Method. It is utilized to work with a creep
model for a constant elastic strain. The creep strain at each cross section can then be
calculated, adjusted and stored at each time increment phase, load changes and
deflection modifications.
5
3. Proposed a theoretical mechanism based on both test and analysis results to explain
the behavior of stress and strain developments in slender R/C column buckling under
sustained loading.
1.4 Outline of Research
Three equations are established at a column cross section by combining force
equilibrium, constitutive law, and compatibility conditions. The strains and curvatures at
each section of the column are resolved. Knowing boundary conditions and all
curvatures at each cross section, the deflections at each section can be obtained by using
numerical approximate methods.
In the instant loaddeflection analysis process, with projected transformations, a
spatial deflection curve is resolved into a couple of planar curves located separately in
two orthogonal plans. Given coordinates of a point at a cross section, the axial strain can
be calculated with curvatures of the two planar curves and the maximum concrete
compression strain at a corner. Based on force equilibrium equations of inner force at a
column section, a set of three simultaneous nonlinear integral equations are derived to
establish the relationships between the planar curve functions with the eccentric load
upon the top of the column. Utilizing the Green's Integral Formula, the strain and stress
nonlinear relationships and column section properties can be solely integrated into a few
important coefficients of the equations. Thus, the approach is suitable for columns with
nonrectangular sections and any kinds of constitutive laws of materials.
During the sustained loading process, creep and shrinkage strains have occurred
and they are added and combined to the total strain and curvature of each section.
6
The present analysis proposes a computerized method to allow creep increment to be
adjusted with increasing of time and variation of strain, the
Time and Strain
Adjustment of Creep Method (TSACM), combining the creep analisis models used for
constant elastic strains, such as those regulated in American Concrete Institute (ACI)
20992 and the Comite' EuroInternational du Be'ton (CEB)FIP 1990 Model Code. The
creep strain at each cross section can be stored, calculated and adjusted and at each time
increment phase, load changes and deflection modifications.
The present rational computer analysis results have been compared with the
existing biaxial and uniaxial experimental data available in the literature. It is indicated
that the results from the proposed analysis correlate with experimental data well.
Some important material and geometry variables effects on the behavior of
slender columns under sustained loads are analyzed. Results are compared with those per
ACI 31899 building code.
In this study, C Language is utilized to compose source code.
CHAPTER 2
LITERATURE REVIEW
2.1 Classification of Concrete Timedependent Properties
Timedependent properties of concrete have played an important part in the buckling of
slender R/C columns under sustained loads. The timedependent behavior of concrete
members is normally classified as creep and shrinkage.
Creep is defined as the timedependent increase of strain in hardened concrete
subjected to sustained stress (ACI 209, 1994).
Shrinkage is defined as the decrease in concrete volume with time after hardening
of concrete. That decrease is due to changes in the moisture content of the concrete and
physicalchemical changes, which occur without stress attributable to actions external to
the concrete (ACI 209, 1994). As the shrinkage is not related to loads, analysis results
show that it is far less as important as creep to affect the behavior of slender columns
(Mauch, 1966), which is to be confirmed in this study.
Figure 2.1 Changes in Strain of a Loaded and Drying Specimen (Wang and Salmon,
1992)
Figure 2.1 shows the relationships between elastic, creep and shrinkage strains
subjected to sustained loading; t o is the time of application of load.
7
8
2.2 Mechanism of Concrete Creep and Shrinkage
2.2.1 Factors Affecting Creep and Shrinkage
Normally, creep of concrete is classified as two parts: basic creep and dry creep (ACI
209, 1994). Basic creep is also called real creep, which is developed under the condition
of humidity equilibrium. This creep value is measured from the specimens under sealing.
It is related to constant elastic stress and loading duration (Wang et al, 1985). Dry creep
is the result of the exchange between the environment and tested specimen, which is
increased with time (Wang et al, 1985).
Concrete creep is a complicated phenomenon due to a wide variety of factors
affecting it. Those factors are normally recognized as (Han, 1996):
1.Cement type and fineness of cement:
Different types and fineness of cement result in different hydrated rates of cement
paste. At the same concrete age, cement paste with different hydrated rates will achieve
different proportions of the final strength (CEBFIP, 1993). It has been concluded (Han,
1996) that the slower the strength is being developed, the larger the potential of creep
strain will pose. This effect has been considered in CEBFIP Model 90 (1993), for both
calculations of creep and shrinkage. In that model, types of cement are classified as SL,
slowing hardening cement, N or R, normal or rapid hardening cement, and RS, rapid
hardening high strength cements. Refer to Appendix B for more details about cement
classification in CEBFIP Model 90.
2.Aggregates grading and the property of admixtures:
Normal weight aggregates, stone and sands, exhibit extremely negligible creep
comparing to cement paste. Thus, good grading aggregates effectively resist the creep
9
tendency of cement paste due to stress redistribution inside of concrete admixture. The
main factors of the resistance are the amount of coarse aggregate and the modulus of
elasticity of the aggregate (Han, 1996). ACI209R92 (1994) includes this factor by
introducing fine aggregate percentage correction factor The larger the fine aggregate
percentage is, the larger the creep will be. Refer to Appendix A for more details about
correction factor Iv
3.Water cement ratio:
Cement content does not influence concrete creep, but water cement ratio does
(Wang and Salmon, 1992). It is due to the fact that increase water cement ratio that also
will increase free water in concrete admixture. This may firstly increase the amount of
the micro cracks developed in concrete during the hydrated process; secondly decrease
the cement paste strength, which turns out increasing the possibility of stress
redistribution from cement paste to aggregates under sustained loading, thus increases
concrete creep (Han, 1996). Since water cement ratio affects concrete slump value, ACI
209R92 introduces the creep correction factor of slump 7, to include water cement ratio
effect on creep. Refer to Appendix A for calculation of correction factor γs.
4.
Curing humidity and temperature:
Relative humidity and temperature are the two important factors for concrete
curing. Obviously, better curing will achieve better quality of concrete products, which
means higher density and less micro cracks, therefore less creep. Thus, a certain level of
relative humidity and temperature are helpful to curb the development of concrete creep.
On the other hand, a long duration of high level relative humidity (such as 100%)
environment for curing process before loading, may also store a lot free water inside of
10
concrete, which turns out enhancing creep (Han, 1996). Steamed curing is better than
the moist for concrete to curb development of creep. This has been reflected in the
ACI20992R Model calculation. Refer to Appendix A for detailed calculation.
5.Age of concrete at loading:
The older the age of concrete means the longer the hydrated process endued, thus
less creep potential to develop. Both ACI and CEBFIP models have included this effect.
6.Relative humidity and temperature after loading:
After the loading, higher relative humidity and a certain level of temperature are
effective to decelerate the loss of free water from admixture and promote the developing
hydrated process of concrete. Thus, less creep would develop. Both ACI and CEBFIP
models have included this effect.
7.Concrete stress/strength level:
It is generally accepted that until a stress level, the concrete creep is proportional
to elastic compressive stress (Wang and Salmon, 1992). Above that level, significant
internal microcracks happen and extend widely, thus accelerate creep (Han, 1996).
Under a certain level of concrete compressive stress level, about 40% of ultimate
strength, both ACI and CEB introduce linear product models. Above that level, CEB
uses a nonlinear exponent function to consider high stress effect. See Appendix B for
detailed calculation.
8.Dimension effect:
Theoretically, dimension effect could be zero if no dry contraction (dry creep)
happens (Wang et al, 1985). In another words, if a specimen is sealed in 100% relative
humidity environment from curing until the end of loading, then the specimen size should
11
have no effects on concrete creep. In practical situations, due to different relative
humidity conditions, that effect does exist, and sometimes very large (Wang et al, 1985).
The ratio of volume/surface areas is used to estimate a dimension effect on creep (ACI
209, 1994). The larger the value is, the slower and the less loss of free water will be in
the admixture, thus, the less creep could happen (Han, 1996).
Figure 2.2 Relationships between Creep and Various Factors Which Influence Creep
The relationships between creep and various factors, which influence creep, are
shown in Figure 2.2 (Han, 1996). Generally, those factors are primarily related to
12
moisture loss, which influence creep, have been found to influence shrinkage as well
(Wang and Salmon, 1992).
Although creep and shrinkage are two different chemical and physical processes
of concrete, but development of each process relates the other one closely. For example,
shrinkage produces micro cracks, which might contribute more creep to happen. On the
other hand, under the external loading pores inside the concrete become smaller, thus
resulting a slow down of the transport of free water to the environment. Therefore, the
development of shrinkage could be slower.
2.2.2 Mechanism of Concrete Creep
1. Viscous Flow Theory
In this theory, cement paste is assumed as to be a high viscous fluid, which
increases viscosity with time as development of chemical changes continues. When
concrete is loaded, the viscous paste is flowing under compressive stress. The movement
will gradually transfer the stress on cement paste to the aggregate. The rate of creep will
progressively reduced as the stress redistribution ends. However, viscous flow theory
cannot explain the phenomena of volume changes within the creep process. Further
more, the theory also assumes proportionality between stress/strain and rate of stress over
rate of strain. It is suitable for concrete compressive stress is in the range of 30 to 50% of
the strength. That stress upper limitation is believed due to the development of micro
cracks. Nevertheless, the viscous flow theory can be used to describe mechanism of
creep with some observations (Neville, Dilger and Brooks, 1983) (Han, 1996).
13
2.Plastic Flow Theory
The plastic flow theory is similar to the one that explains the mechanism of metal
creep. It suggests that creep of concrete is caused by a plastic flow called crystalline, i.e.
a slipping along planes within a crystal lattice. The plastic flow is a result of a slip in a
plan of maximum shear stress. The water inside concrete is assumed to act as lubricant
thus makes flow easier. One of the shortcomings of this theory is, for metals at room
temperatures, the plastic flow occurs only when the applied stress equals or exceeds the
yield point. While in the case of concrete, there is no such a yield point. Additionally,
the theory cannot explain volume changes either. However, at very high stress level, the
theory seems explaining concrete creep better (Han, 1996).
3.Seepage Theory
In this theory, the creep of concrete is explained as seepage of gel water under
pressure. The gel water inside of concrete admixture is in a state of equilibrium under
vapor pressure. When an external load is applied to the concrete, the pressure is changed
and equilibrium broken. The gel water is reorganized to obtain the new equilibrium. It
affects the colloidal stresses and van der Waals force. A new equilibrium is reached by
seepage of gel water inside the cement gel. It is obvious that the density (porous) of the
cement paste has a strong effect on the seepage. One of challenges to this theory is how
to explain the remained creep in the unloading process. According to the theory, after
removal of the external load, due to reducing of pressure of gel water, the original
equilibrium would reestablish. The result of creep recover test cannot support this point.
One basic argument about the seepage theory is that, if the gel water is slowly squeezed
out of the pores and the capillary water evaporates, a considerable weightless could be
14
happened during the creep process. Nevertheless, this could not be proved by experiment
either (Neville, Dilger and Brooks, 1983) (Han, 1996).
4. Microcracking Theory
The development of microcracks associated with creep has been confirmed by
results from acoustic measurement. It is estimated that micro cracking is taking 10 to
25% of the total creep deformation in concrete. The extent of the development of micro
cracking being due to creep depends on existing microcracks prior to loading. At high
stress level, micro cracking contributes considerably to concrete creep (Han, 1996).
Due to its complicity, even though quite a few research results have been
achieved on concrete creep, to clarify the mechanism of concrete creep is still very
difficult. As being described by Illston et al. (1979): " The physical and chemical
happenings that are associated with creep are on a molecular scale, and there is no
convincing direct evidence of what actually goes on; so the explanation of creep has,
perforce, consisted of interpreting engineering level observations in terms of likely
physical and chemical phenomena." It is generally agreed none of the mechanisms
proposed up to date can explain all of observed facts.
2.2.3 Mechanism of Concrete Shrinkage
Shrinkage of concrete is classified as plastic shrinkage, chemical shrinkage and drying
shrinkage (Gilbert, 2001).
Plastic shrinkage, which happens in the wet concrete, may result in significant
cracking during a setting process. This cracking occurs due to capillary tension in pore
water. Since the bond between plastic concrete and reinforcement has not yet developed,
15
the steel is ineffective in controlling such cracks. This problem may be severe in the case
of low water content and existing silica fume concrete. Thus, high strength concretes are
prone to plastic shrinkage (Gilbert, 2001).
Drying shrinkage is caused primarily by the loss of water during the drying
process. Chemical (or endogenous) shrinkage results from various chemical reactions
within the cement paste and includes hydration shrinkage, which is related to the degree
of hydration of the binder in a sealed specimen (Gilbert, 2001).
Concrete shrinkage strain is usually considered as the sum of drying and chemical
shrinkage components (Gilbert, 2001).
2.2.4 Different Properties of Creep and Shrinkage between NSC and HSC
HSC is not only characterized by superior engineering properties such as higher
compressive strength, elastic modulus and density, but also by remarkably different
material physical properties such as more discontinuous and closed pore structure,
smaller pore sizes, and a more uniform pore distribution. This , among others, is a result
of addition of silica fume or the like. Those differences in the microstructure cause
changes in properties like creep and shrinkage. (Han, 1996)
It has been found that creep is much lower in HSC than NSC, especially under the
drying conditions. The ratio of sustained loading strength to shortterm strength is also
higher in HSC than NSC. Results from experiments have confirmed that the mount of
cracks including those related at initial state is much less in HSC. Consequently, smaller
inelastic deformations and higher creepstress linearity limits have also been observed
(Han,1996). In CEBFIP Model 90, concrete strength has been considered in calculating
16
the creep and shrinkage coefficients (Refer to Appendix B for more details). The higher
is the strength, the smaller are the values of those coefficients.
2.3 Analytical Models to Predict Creep and Shrinkage
There are three types of models which have been developed to predict concrete creep and
shrinkage. One method is a material science approach. In that approach, the models are
developed on theories about microstructure of concrete material. As pointed out by
Wittmann (1982), the models are very complicated in presentation and therefore very
difficult for use in practical engineering. Another method is to formulate empirical linear
models on basis of test data. Although it still suffers some basic theoretical weakness, it
is a more accrurate method in most practical situations. Particularly, it is much easier to
be adopted in engineering application. The third approach is to create a nonlinear creep
model based on rheology theory (Shen, 1992), (Walraven and Shen, 1993). Compared
with microstructure model, it has avoided using the complicacity of microstructural
theory and exponent presentation. However, most parameters in the formula are still
needed to be determined from the experiments.
2.3.1 ACI 209R92 Models
1. Creep Model
A general empirical method, developed by Branson (1971), which has been
accepted by ACI209, gives a creep coefficient equation under loading ages of 7 days for
moist cured concrete and 13 days for steam cured concrete:
where t
is the duration of loading (days) and 14, is the ultimate creep coefficient.
γc denotes the product of the applicable correction factors defined in Eq. (2.3)
hi, denotes the correction factor for loading age
72
denotes the correction factor for ambient relative humidity
rat denotes the correction factor for average thickness of member other than 6 inch
γs
denotes the correction factor for slump
iv denotes the correction factor for fine aggregate percentage
γa denotes the correction factor for air content
See Appendix A for calculations of the factors. In this research, since all test data
collected do not provide air content information. An assumption of 6% is made for the
creep estimation of all concrete specimens.
Figure 2.3 shows typical creep variation with duration of loading (Wang and
Salmon, 1992).
17
Figure 2.3
Typical Relationships between Deformations of Elastic, Creep and the
Recoveries
2. Shrinkage Model
Shrinkage after age 7 days for moist cured concrete is given by:
where t is the time after shrinkage is considered, and (εsh)u
is the ultimate shrinkage
strain.
Shrinkage after age 13 days for steamcured concrete is given by:
In absence of specific shrinkage data for local aggregates and conditions, the
average values suggested for (εsh)
u are:
where,
γsh denotes the product of the applicable correction factors defined in Eq. (2.7)
γcp denotes the correction factor for initial moist curing
γc denotes the correction factor for cement content
See Appendix A for calculations of factors yip
, 72, nit, γs, γc, and
Figure 2.4 shows a typical shrinkage strain variation with time after moist curing
(Wang and Salmon, 1992).
18
Figure 2.4 Standard Shrinkage Strain Variation with Time after Moist Curing
3.
Concrete compressive strength versus time
where, a in days and
/3
are constants, ( f')28=28day
strength and t
in days is the age of
concrete. The values of a and pi
are related to the type of cement used and the type of
curing employed. Typical values recommended are given in Table 2.2.1 of ACI 209R92
(ACI Committee 209, 1994).
2.3.2 Models of CEBFIP Model Code 1990
1. Creep Model:
When stress ac
<0.4f cm(to),
the creep is assumed to be linearly related to the
stress. For constant stress applied at time to this leads to:
where,
fcm(t0) is the mean concrete compressive strength at an age of days to
E ci is the modulus of elasticity at the age of 28 days
20
0 (t, t0) is the creep coefficient. It may be calculated from:
00
is the notional creep coefficient, refer to Appendix B for more information.
βc(tt0) is the function to describe the development of creep with time after loading (see
Appendix B for more information)
t is t he age of concrete at the moment considered (in days)
t0 is the age of concrete at loading (in days)
For a stress level in the range of 0.4fcm(t0)< σc<0.6fcm(t0)
the nonlinearity of creep may
be taken into account using the following Equations:
where,
00.k
is the nonlinear notional creep coefficient, which replaces O0
in Eq. (2.10)
2. Development of strength with time
For a mean temperature of 20°C and curing in accordance with ISO 2736/2, the
relative compressive strength of concrete at various ages fcm(t)
may be estimated from
Equations (2.14):
is the mean compressive strength of concrete at an age of 28 days
21
fcm(t) is the mean compressive strength of concrete at an age of
t days
fl (t) is a coefficient, which depends on the age of concrete t days
s is a coefficient which depends on the type of cement: s=0.20 for rapid hardening high
strength cements RS, 0.25 for normal and rapid hardening cements N and R, and 0.38 for
slowly hardening cements SL.
t 1 equals 1 day
3.Strength under sustained loads
The combined effect of sustained stresses and of continued hydration is given by
Equations (2.16) and (2.17)
with
where,
fcm,sus(t,to)
is the mean compressive strength of concrete at time t when subjected to a
highsustained compressive stress at an age at loading to<t
βc,sus(t,to) is a coefficient which depends on the time under high sustained loads tto
(days). The coefficient describes the decrease of strength with time under load and is
defined for (tt0)>0.015days (=20 min)
4.Shrinkage Model
The total shrinkage strain εcs(t,to) may be calculated from:
εcs0 is the notional shrinkage coefficient
22
/3 is the coefficient to describe shrinkage with time
t is the age of concrete (days)
t s is the age of concrete (days) at the beginning of shrinkage
Refer to Appendix B for more details of calculations.
2.3.3 Nonlinear Model for Creep
According to Walraven and Shen (1993), the nonlinear model contains two nonlinear
springs and a nonlinear dashpot (see Figure 2.5).
Figure 2.5 A Nonlinear Model for Concrete Consisting of Two Springs and One
Dashpot [after Shen (1993)]
A general constitutive equation can be derived in the following:
where,
w is the creep rate factor
ck is the nonlinearity factor for spring Hk
23
θk is the strain level, in. Hk i.e. εk/εku
εku is the ultimate strain of spring
Hk
am is the stress level, i.e. σm/fm
.1=0, for creep
to is the concrete age at loading
The characteristics of the spring H can be described in the following:
where,
is the nonlinearity factor for spring Hm
is the strain level, in. H
m
i.e.
en /Emu
is the material factor for unloading
is the ultimate strain of spring Hm
is the highest stress level in the history
2.3.4 Microstructure Model for Creep
Wittmann et al. (1982) used the activation energy approach to describe the creep of
concrete. It is assumed that the movement of solid particles is responsible for the creep
of concrete.
24
where,
εcr is the creep of concrete
A 1,A2 and 113 are constants
V, is activation volume
Qo is a constant
R is gas constant
T is absolute temperature
a is the applied stress
me is a constant
t is the concrete age
2.4 Studies on Slender RC Column Buckling under Sustained Loads
2.4.1 Effects of Creep and Shrinkage on Strength of RC Column
Except for noneccentrically loaded columns, any occurrence of creep will increase
geometrically the bending curvature of a section in a column, because the higher
compressive strain yields the higher creep strain. As a result, the increasing of bending
curvatures of column sections yields the increasing of the lateral deflection of the
column, which in turn increases the axial load's eccentricities to column sections thus
affecting the internal concrete and reinforcing steel strains and stresses. Although the
development of creep may help redistribute part of concrete compression stress to steel
reinforcing in compression zone, concrete creep might still decrease the buckling
capacity of slender columns.
25
As it is not caused by concrete stress changes, shrinkage will tend to relax the
concrete compression stress and redistribute the force difference to longitudinal
reinforcement. That is, to increase the compressive stress in steel reinforcement
previously in compression, and /or to decrease the tensile stress in steel reinforcement
previously in tension. As a result the moment curvature of a section decreases, this, in
general, is beneficial to the increase of buckling strength of the slender columns.
2.4.2 Experimental Studies
From the mid 1950s, academic researchers started to pay attention to studying the slender
RC columns under sustained axial loads. During 1960 through 1970, the research and
experiments in this field became very active. Breen and Green (1969) from University of
Texas investigated behavior of restrained and unrestrained reinforced concrete columns
under sustained loads in their Ph.D. programs. Goyal (1971) from Scotland, performed
study of (pinned) slender columns under sustained loads. Lately, Gilbert and
Mickleborough (1993) studied the creep effects in slender reinforced and prestressed
concrete columns. When HSC is widely used in engineering practice, some researchers
began their focus on the experiments with the R/C column made of HSC (Claeson, 2000).
Until now, most test data are still from the specimens made of NSC.
Drysdale et al. (1971) published their research on sustained biaxial load on
slender concrete columns. This is the only published paper that has focused on the
experiments in biaxially loaded slender RC columns under sustained loads. Due to the
loading equipment and other conditions limits, the scales of test specimens are about 1:5
1:2; the sustained load durations are less than 3 years; all specimens are pinned top and
26
bottom and loaded eccentrically and symmetrically. Per ACI129R92 model, concrete
could have reached 87% of the total creep strain in 3 years. Per CEBFIP 90 model, that
number becomes 7790%. Thus, most of creep has developed in that period, which
means, the 3 year experimental results should have reflected the real creep process from
this point of view. Refer to Chapter 4 for more discussions of the experimental results.
2.4.3 Theoretical Analysis
Ostlund (1957) suggested the Reduced Modulus of Elasticity (RME)
to estimate the
effect of creep on buckling load of RC column. According to Ostlund, the reduced
modulus of elasticity of concrete E r can be obtained from the following formula:
where,
a denotes liquidity constant in the creep law for the method of superposition
y
o denotes liquidity constant in the creep law for the rate of creep method
a denotes time constant in creep law
E denotes concrete elastic modulus
Ostlund's study opened the theoretical analysis in this field. The advantage of the
method is its simplicity and easy to be applied in engineering. In fact, this concept has
provided the basic direction for the ACI318 Building Codes since 1971.
The difficulty with this approach is that the value of
E/Er is not directly known
and varies in a large arrange. In ACI318, (1+
βd) is used instead of
(1+αE)
to get the
27
reduced
El
value. It has neglected the value variations of concrete creep. In particular,
the reduced modulus approach could not be used reliably to determine the resistance of a
column to a long duration (dead) load, followed by an additional (live) load. (Mauch,
1966)
Bresler and Selna (1964) proposed a powerful method for creep buckling of RC
slender column analysis. It is called the Relaxation Procedure. In this method, during a
time interval, the strain distribution is initially held unchanged. Under the situation, if the
creep and shrinkage components change, the elastic components of total strain must also
change by the same amount in the opposite direction. The internal force equilibrium
could not be kept unless both a change of axial force AN and a change of bending
moment AM are applied to this section. The strain state is thawed (relaxed) after the
changes of internal forces being applied to the section. This approach is accurate for the
material that is linear elastic at shortterm.
That is, at section i,
where,Δε0i
is the change in strain of the top fiberΔki
is the change in curvature of the section
A=fdA is the area of the transformed sectionB=∫ydA
is the first moment of the transformed area about the top surface of the section
28
I is the secondary moment of the transformed area about the top surface of the section
E,
is the concrete elastic modulus
AM, is the change of bending moment
AN, is the change of axial force
Bazant et al. (1972) proposed the Ageadjusted Effective Modulus Method
(AEMM) to model the effects of creep.
where,
E" (t,t0) is the ageadjusted effective modulus
E(t0) is the instantaneous elastic modulus in time t0
Φ(t,to) is the increment of the timedependent concrete creep coefficient associated with
the time interval and load increment under consideration
X(t,to ) is the aging coefficient for concrete during the same interval
Combining Bazant's AEMM and Relaxation Procedure proposed by Bresler and
Selna (1964). Gilbert et al. (1993) proposed a nonlinear procedure for the time
dependent analysis of reinforced and prestressed concrete columns under uniaxial
sustained eccentric compression. In their method, both material and geometric non
linearites are taken into account in an iterative computerbased solution procedure, in
which, for shortterm analysis, stressstrain relation for concrete is bilinear with crack
point in between. Individual crosssections are analyzed using the ageadjusted effective
modulus method to include the effects of creep and shrinkage. In which,
29
where,
Aso, is the change in strain of the top fiber
AK, is the change in curvature of the section
A, is the effective area of the transformed section
Be is the first moment of the effective transformed area about the top surface of the
section
I, is the secondary moment of the effective transformed area about the top surface of the
section
AM, is change of bending moment
AN, is change of axial force
Dividing the time scale into several increments, the gradual development of time
dependent cracking can be traced as the lateral deflection of the column and the internal
secondary moments increase with time due to creep. In the numerical analysis part, the
variation of curvature along the member is assumed to be parabolic. That is:
where,
6 is the lateral deflection of column at midheight
kA is the curvature of top section
30
KB is the curvature of bottom sectionkC
is the curvature at midheight section
L the length of column
Claeson et al. (2000) presented the analysis of slender HSC columns under uni
axial sustained eccentric loads by using CEBFIP 90 Model Code (1993). The analysis
used a momentcurvature approach. The calculations were based on a crosssectional
analysis. The second order effects were taken into account by assuming a sine curve
deflection shape. In addition, the confining effect of the stirrups was carried out. Creep
of the concrete was taken into account using the Equation (2.8) from CEBFIP.
The modulus of elasticity for concrete was not assumed to increase with respect to
time. The experiment pointed out that, HSC columns exhibited fewer tendencies to creep
and could sustain the axial load without much increase in deformation for a longer
period. They also exhibited less nonlinear creep.
For the RC slender columns under sustained biaxial bending loads, Drysdale et al
presented an analysis method in 1971. To describe twodimensional distribution of strain
on sections, cross section was divided into a grid. Using the symmetric feature for
loading at top of a pinned column, only half was analyzed. The half column was divided
into four segments. The distributions of strain on sections were assumed first manually
and then corrected by computerized iteration process controlled by three internal force
equilibrium equations at each section and the moment eccentricity at the end of column.
For creep prediction, a nonlinear creep model was used and the
Modified Superposition
Method for Creep Prediction
was adapted to this computerized analysis.
31
Figure 2.6 Modified Superposition Method for Calculating Concrete Creep (Drysdale,
1967)
From Figure 2.6 the total creep at t2 equals:
where,
creep]: creep strain under initial constant elastic strain EL,1 during time to to ti
creep2: creep strain under initial constant elastic strain EL.2 during time t1 to t2
and
Cc EL,2
during time to to t2ticreep2": creep strain difference between lines of cc EL,
Cc EL,1: creep strain line under constant stress σ1
Cc EL,2:
creep strain line under constant stress 0
2
32
Obviously, the shortcoming of this method is that the concrete creep will be over
estimated when stress is decreasing, which frequently occurs at early phase of sustained
loading.
Mauch (1965) also performed theoretical study on RC slender column creep
buckling based on rheology theory, which resulted in a very complicated mathematic
formula. According to his research, slender column reductions in carrying capacity due
to creep were 40 to 50 percent for columns with a slenderness of L/D=25 to 40, where L
is the length of column and D is the diameter of column. Smaller reductions occurred for
shorter columns.
2.5 Engineering Approach Specified by Engineering Design Codes
In 1963, ACI made substantial modifications to the design procedures applicable to
slender columns, including the effects of sustained loads in some cases. The Code
required that for a certain type of restrained column with L/D>30, an analysis taking into
account the effect of additional deflections on moment was required. The Code retained
the use of a long column reduction factor. The reduction factor was expressed as a linear
function of the slenderness ratio and applied equally to load and moment (ACI318,
1963).
In 1971, ACI introduced the Moment Magnifier Design Method for design of
slender columns under both short term and sustained loads. The design method has been
adopted till today.
33
1. ACI 31899 Provisions:
In ACI 31899 Building Code, the effect of sustained load on slender column is
considered in calculation of Stiffness Parameter EL which is taken as following:
where,
E, denotes concrete modulus of elasticity
/g denotes gross moment of inertia of concrete section
lid
denotes proportion of the factored axial load that is considered sustained
factored sustained axial load
factored total axial load
Es denotes steel modulus of elasticity
Is denotes gross moment of inertia of reinforcement
The value of El
is needed for calculating the Euler buckling load P e of a column,
which affects the strength of a RC slender column.
where,
kL, is the effective length of column
In case of a braced slender column, the design moment Mm
is:
34
where,
M2b is the larger end moment acting on the member
M1b is the smaller end moment acting on the member
M2b/ MTh
positive for single curvature
Mm is primary bending moment due to transverse loading
Cm is a factor in Moment Magnifier for braced Frames
End moments only:
where,
Pu
is a factored axial force
Here the effect of the sustained load directly reflects on the value of β
d. Ai is a
component of the column section stiffness parameter
El. El affects the Euler buckling
capacity of column
Pc,
which influences the value of magnifier
81)•
MacGregor et al (1970) indicated the reason to choose
d,
(Rm
noted at that time)
for creep consideration:" This factor has been chosen to give the correct trend when
compared to analysis and tests (Green, 1966) of columns under sustained loads".
35
Obviously, there are two problems about the Moment Magnifier Method. First, it
is not based on any theories of concrete timedependent behavior. Second, it needs to
analyze more test data results to improve the method.
2. CEBFIP Model Code 1990:
In CEBFIP 90, the creep effects have been introduced as the creep eccentricity ec
where,
e0 denotes the first order eccentricity
ea denotes the additional eccentricity introduced by the effects of geometrical
imperfections
Nsg
denotes the axial force in the element, under the quasipermanent combination of
actions
NE
denotes the critical Eulerload of the column.
In this model, the creep coefficient Φ(t,t0)
(see Eq.(2.9)) has been included, which reflects
the quantity level of creep influence.
2.6 Summary
1. Results of Previous Studies
a).
Creep decreases the carrying capacity of slender reinforced concrete columns.
b).
HSC columns exhibit fewer tendencies to creep and can sustain the axial load and
exhibit less nonlinear creep.
36
2. Unsolved problems
a).
There are no experimental test data available for biaxial bending slender HSC
columns. Only one experimental work is known for NSC columns under biaxial
sustained loads.
b).Due to the conditional limit, the experimental longest duration of sustained loads is
less than 3 years, which means a possible 1520% creep effects have not been studied.
c).In theoretical analysis, the calculation of creep due to stress changes has not been
thoroughly studied.
d).Most analysis research were performed in 1960s, the models for creep and shrinkage
are needed to be updated.
e). Unlike CEBFIP 90 Models, the design method in the ACI building codes is not
established on the studies of concrete creep and shrinkage. Consequently, the current ACI
building code is based on the results on concrete slender columns under sustained loads
around 1970.
CHAPTER 3
PROPOSED ANALYTICAL APPROACH
A new numerical method is developed to study the behavior of biaxially loaded slender
RC columns, made of NSC/HSC, under sustained loads. In this approach, both material
nonlinearity (concrete) and geometry nonlinearity are considered. Latest concrete time
dependent models are adopted. It is also a generalized method in terms of loading
history, column supports, material (concrete) stressstrain behavior and shapes of column
sections (unchanged through length of column). Except for basic dimension, loading
history, support condition, and material properties, no further manually initial input data
are needed.
3.1 Analysis Strategies
As it is a biaxial bending problem, all forces, (axial force, moments and stresses) and
deformation variables (strains, curvatures, and deflections) are projected to
XZ and YZ
planes separately. For example, the curvatures on a section 0 are divided into c,
and Φy
respectively,
The solution process is based on an iteration method:
1.The columns are divided into a number of segments. At start end of each
segment, a cross section is recorded to present the segment on which three
simultaneous differential equations of internal force equilibrium are established.
2.For the shortterm loading process, the concrete maximum compressive strain ece
and two orthogonal projected curvatures 0, and of a section can be solved from
37
38
the three simultaneous differential equations on a cross section. The initial
iteration data is obtained from linear elastic analysis.
Figure 3.1 Projected Curvatures of
O
and Φy
3.Knowing the curvatures of all segments, the deflection line of the column can be
obtained accurate enough with a numerical method.
4.
To include the effects of second order deflection, a second round of iteration is
performed. The control of the iteration is the error allowance between the two
consecutive rounds of maximum deflection of the column.
5. For sustained load duration, the loading period is divided into a series tiny time
intervals. At each interval, the creep and shrinkage strain components are
calculated by concrete creep and shrinkage theoretical models and adjusted
according to the TSACM,
the method generated herein. The key point is that,
TSACM
is able to adjust creep under the situation while concrete elastic stress is
decreasing, which is a typical phenomenon for reinforced concrete column creep
buckling. Concrete strength and modulus of elasticity are age adjusted at each
time interval.
39
3.2 Analysis Assumptions
The following basic assumptions are made for the analysis:
1.
Linear distribution of strain across the section is valid.
2.Concrete is treated as continuous uniformed medium.
3.No tensile stress is considered for concrete.
4.
Bond between concrete and reinforcement is perfect.
5. Under sustained loading, the age and /or stress adjusted concrete stressstrain
equation of short term is valid for describing the elastic stressstrain changes.
3.3 Solution Procedures
3.3.1 Basic Equations
1. Equilibrium Equations
According to the coordinate system shown in Figure 3.2, the force and moments
equilibrium equations on a cross section are:
Figure 3.2
System of Global Forces and Deflections
Substitute symbols of external and internal forces of into the above equations, thus:
where,
P denotes axial load
Moy
denotes Moment at top of column rotating over y axis
Mox denotes Moment at top of column rotating over x axis
Nz[σz(x,y)] denotes stress resultant along z axis
My [σz(x,y)] denotes stress moment resultant rotating over y axis
Mx[σ(x,y)] denotes stress moment resultant rotating overx axis
u(z)
denotes deflection at the center of the section projected along x axis
v(z)
denotes deflection at the center of the section projected along y axis
σz(x,y) denotes concrete compressive stress at point (x,y,z)
x, y, z denote global coordinates
2. Constitutive laws:
Since tension in concrete is neglected, only compression is considered. For both
NSC and HSC, the Modified Hognestad Formula (Park, et al, 1975) (Kent, et al., 1971) is
applied:
40
a is the concrete compressive stress
s is the concrete compressive strain
f' is the concrete cylinder strength
s0 is the concrete strain refers to maximum strength, take 0.002
εu is the concrete ultimate strain, use 0.003
6.50 , 550u
and
550h,
refer to Figure 3.3 for physical meanings
b" is the width of confined core measured to outside of hoops
sh is the spacing of hoops
ps is the reinforcement ratio
42
Figure 3.3
Concrete Compressive Stressstrain Relationships (Park and Paulay, 1975)
4.
Stress and Strain relationship of steel reinforcement:
where,
as is the stress in steel reinforcement
es is the strain in steel reinforcement
E, is the modulus of steel
fy is the steel yield point
4. Compatibility Equations:
43
where,
s
is the strain of a point (x,y,z) at the section z
ec0
is the strain of the point (0,0,z) at the section
z
Ox
is the projected bending curvature on plan XOZ
is the projected bending curvature on plan YOZ
5. Time Dependent Models:
Use ACI12992R Models for NSC analysis, and CEBFIP 90 Models for HSC.
For concrete strength changing versus time under sustained loads, use CEBFIP model.
3.3.2 Iteration Controlled Derivative Equations
The internal forces can be represented by stresses from the balance of the internal
stresses:
where,
D refers to the domain of concrete compression.
44
denotes the reinforcement area at j point
yj denote the global coordinates at j point
Stresses can be represented further by strains from the basic concrete and steel
constitutive equations:
reorganize:
Introducing:
Since
where,
εec0, Φxand Φyare constants
Therefore
Derivate both sides of Eq (3.28):
45
Thus the first term of right side of Eq (3.25) becomes:
Notice Green's Integral Formula:
In that L represents the line envelop of the region D. And the line integral goes with
anticlock direction.
For rectangular sections, Eq. (3.37) can be changed to:
Substitute Eq. (3.39) into Eq. (3.38):
Bring results of Eq. (3.40) into Eq. (3.25), therefore:
Using the same method, Eq.(3.26) can also be modified to the following:
Derivate both sides of Eq (3.29), i. e.
Substitute Eq. (3.30) into Eq.(3.42) and reorganize it :
Bring result of
Eq. (3.43) into the second term at right side of Eq. (3.26):
Applying Green's Formula, the first part at right side of Eq. (3.44) becomes:
the right side of Eq. (3.45) can be changed into:
46
while the second part at right side of Eq. (3.44) becomes:
For rectangular section, right side of Eq. (3.48) becomes:
47
Then Eq. (3.26) becomes:
48
These three equations are the core equations for iterative solutions. The curvatures Ox,
and boundary conditions, the solution of deflection along column length could be
obtained. Refers to Appendix D for details.
3.3.3 Time and Strain Adjustment of Creep Method
Assuming t0 is the age of concrete when instant eccentric axial load is loaded and
T is the
duration of the sustained load, divide T
into n numbers of minor phases, (1, 2, 3, ...,
n1,n), and denote t, as the age of concrete at the end of i phases. Thus, at a age
t„
the
49
strain of the top fiber in compression stress on a cross section, which, in fact is at a corner
point, would be:
where,
ε(t0,ti) denotes the total strain at top fiber in end of t; phase
t0 is the age of concrete while loaded
Δεe(ti) is the change of elastic strain happened at
ti phase
v(tj,ti) is the coefficient of concrete creep strain developed from tj phase to ti phase
εs'h (to , ) is the effective shrinkage strain developed at end of
ti phase
εsh(t0,ti) is the concrete shrinkage without reinforcement from t 0 to ti
A, is the cross section concrete area
As
is the reinforcement total area of cross section
E, is the concrete elastic modulus
Es is the reinforcement elastic modulus
Let Δεe(t0) denotes εe(t0):
50
3.3.4 Constant Step of Creep Increasing Technique
For reasons of stability and convergence of iteration process, it is preferred to be able to
control the increment step of creep strain constant. Thus:
where,
51
C denotes constant,
i denotes number of steps
since il l (x) is nonlinear and close to log function, thus when i→∞,
3.3.5 Main Algorithm Flow Chart
The main flow chart of the algorithm is shown in Figure 3.4. Refer to Appendix F for the
information of a detailed algorithm flow chart.
Figure 3.4 Main Flow Chart of the Algorithm
3.4 Convergence and Stability of Solution
To solve the equilibrium equations of force and moment on a section (3.53), (3.54), and
expressed in the following:
t) is the vector of external force on the top of column,
s
the vector of lateral deflection at the center of the section concerned,
the vector function of iteration.
The equations for convergence are:
52
Under each grade of load actions
Pi(e
x,e
y,t), the convergence of the lateral
deflection curves controls the second round iteration cycle procedure. The k+1 th round
of iteration solution of deflection vector can be expressed as follows:
53
3.5 Limitations of Proposed Analytical Model
In this study, there are some limitations to the application of the proposed analytical
model due to the following reasons:
1.The constitutive law of concrete used in this model is originally generated from the
study of unconfined NSC by Hognested (Wang et al., 1985), and modified by Park et
al. (1975) in research of confined NSC plus some specimens of HSC. It is one of
many approaches to establish a unified formula for both NSC and HSC in this field.
Since there is a big difference in the strain and stress relationships between NSC and
HSC, this formula modified by Park et al. (1975) might not work quite well with HSC
as it with NSC. Therefore, for the analysis of columns made of HSC, further studies
are still needed to study the suitability of Park et al's formula and other constitutive
laws of concrete describe HSC.
2.
As the ACI 20992R models have not included the studies of time dependent
properties of HSC, thus the CEBFIP mode code 90 is used here to estimate the creep
ultimate creep coefficient for HSC columns. Same efforts are made for considering
the creep effects on the concrete with high compressive stress level and the
development of strength with time under sustained loads (see Chapter 2.3.2 for more
details). Since the models of ACI and CEBFIP are established separately, it should
be very careful when using them in some situations (see Chapter 4.3 for more
discussions about the selection of the ultimate creep coefficient value).
3. During the derivation process (Green integral transformation), the concrete is
assumed as continues uniformed medium. It is a simplified approach to the practical
situation. The practical situation is that the concrete in core is confined by stirrups
54
while the rest outside the stirrups is unconfined. As the concrete constitutive laws for
the two different concrete areas are different, it needs special judgment before an
analysis to assume one equivalent concrete constitutive law representing for the entire
cross section. An alternative approach is to make Green integral transformation to
the two concrete zones with different concrete constitutive laws separately. Then, a
new set of derivative equations would need to be set up.
4.
In this model, the creep of reinforcement or reinforcement relaxation has been
neglected. This is due to the facts that, first, for the normal reinforcement at room
temperature the reinforcement of creep or relaxation is much lower as compared with
concrete creep (less than 4.5% for high stress leveled (0.9 6y ) reinforcement
relaxation in 1000 hours to 50% creep for concrete during the same time). Second,
the development of creep or relaxation of reinforcement is much faster than that of
creep (Wang et al., 1985). Third, during most of sustained period the reinforcement
stress level in a slender column is not at high level until the column fails. Since low
level of stress develops low level of creep or relaxation (Wang et al., 1985), the creep
or relaxation of reinforcement could be neglected. As for the situation of using high
strength reinforcement and or sustained high stress level in reinforcement such as
situation in prestressed concrete, the relaxation of reinforcement might need to be
considered.
5.The Proposed analytical model experiences convergence difficulty when the lateral
displacement of a slender column becomes very large.
CHAPTER 4
TEST DATA VERIFICATION
4.1 Slender Column under Sustained Uniaxial Bending Loads (NSC, HSC)
4.1.1 Test Data from Goyal et al. (NSC)
Goyal et al. (1971) performed a total of 46 slender column specimens, 20 of which were
subjected uniaxial bending under sustained loads, while other specimens were short
term loaded. See Figure 4.1, Tables 4.1 & 4.2, for concrete column details, concrete
aggregate grading and ultimate loads information.
Figure 4.1 Reinforcement Details of Goyal et al. (1971)
The specimens were casted and cured at a temperature of 20±1°C with a relative
humidity of 4560% and tested at 28 days of age.
55
Table 4.1
Aggregate Grading of Goyal et al.'s Experiment (1971)
Table 4.2 Column Details and Ultimate Loads
Table 41 Column Detais and Ultimate Loads
56
Creep tests were also carried out in the same environmental conditions. The creep
test results are shown in Figure 4.2. Where, y = creep strain/ initial elastic strain, with a
maximum value of 2.4 at end of 5000 hours of sustained loading duration.
57
Figure 4.2 Variation of y with Time at Different Stress Levels (Goyal et al., 1971)
Specimens R, H & G are subjected to uiaxial sustained loading for 5000 Hours
(208.333 days) and failed under additional instant load after that. The eccentricity is 0.5
inch at top and bottom. From Table 4.1, the predicted shortterm capacity for R is 7.5
kips, for G 12.45 kips and for H 11.92 kips. The constant sustained load for R equals 4.5
kips, for G 7.5 kips and for H 5.0 kips. Thus, the ratio of sustained load/short term load
capacity for R, 0.6; for G, 0.6 and for H, 0.42.
From the companion creep test (Figure 4.2), the ultimate creep coefficient can be
derived as:
Substitute t=208,
2.40 into Eq.(4.1):
vu=3.375
Figure 4.3a Deflectiontime Curve of Column H
58
Figure 4.3b Deflectiontime Curve of Column G
59
Figure 4.3c Deflectiontime Curve of Column R
From Figures 4.3a, b & c, the results of analysis using ACI209 Model with
vu=3.375, conform with the test data well. Except for deflection values of Column G at
5000 hours, the data of test are always larger than those of analysis. It might imply a
modified factor to compensate the differences between test and analysis of all columns.
4.1.2 Test Data from Drysdale et al. (NSC)
Drysdale et al. (1971) completed their experiments of sustained uniaxial and biaxial
loads on slender concrete columns in the Department of Civil Engineering, University of
Toronto. Of the test, eight specimens were subjected to sustained uniaxial bending loads.
See Section 4.2. for more information about the experimental program.
D2A is a specimen subjected to uniaxial sustained loading for 83 days until
failure. The eccentricity is 1in at X axis. From Table 41, the shortterm test capacity is
60
39.9 kips, the constant sustained load equals 30 kips. Thus, ratio of sustained load/load
capacity = 0.75
Per ACI209, one has: vu=1.98, rsh=7.02x10e4
Per CEBFIP 90 Model: 0,
4 =3.10, εcs0=6.47x10e4
Per creep test (see Section 4.2): vu=3.51
From Figures 4.4a & b, the results of analysis using ACI209 92 Model with
vu
=3.62 agree with the data except the results of deflection curves at 80 days of loading.
Since the column fails at 83 days, the difference between test and analysis becomes larger
at 80 days of loading.
Figure 4.4a Deflection Results of D2A
Figure 4.4b Strain Diagrams of D2A (strain unit: 1x10 4)
4.1.3 Test Data from Claeson et al. (HSC)
Claeson and Gylltoft (2000) performed 6 slender columns subjected to uniaxial bending
load. One of the specimens was made of HSC and subjected to sustained load. Figures
4.5 and 4.6 present the details of columns and test equipment. In Table 4.4, fc,cy1 and
fc,cube
refer to compressive strengths on compression tests of sizes 0150x300 mm and
150x150x150 mm, respectively and fc,prism fc,prism refers to compressive strength on compression
test of 200x200x800 specimen.
Deformed bars of 016 used as longitudinal reinforcement, were made of Swedish
Type Ks60s with yield strength 636 MPa, hardening strain 36%o, ultimate strength 721
MPa, ultimate strain 142%0, and modulus of elasticity 221 GPa. While those of 08 used
as stirrups, wree made of Swedish Type Ks40s with yield strength 466, hardening strain
62
22%o, ultimate strength 636 MPa, ultimate strain 87%o, and modulus of elasticity 207
GPa.
Figure 4.5 Details of Columns and Test Equipment in Claeson et al.'s Experiment
(Cleason et al., 2000) (Unit: mm)
Plastic sheeting was used to covere the columns in the first day after pouring. Then the
molds were removed and the columns were cured under the lab room conditions
(18±2°C) and RH=30% for next 27 days. During sustained loading, the environment
conditions were unchanged.
Figure 4.6 Details of Reinforcement in Claeson et al.'s Experiment
Per CEBFIP 90: 00=1.90
Per ACI209: vu=2.06
Table 4.3
Composite of Concrete Mixtures, kg/m3 (Claeson et al., 2000)
63
Table 4.4 Hardened Concrete Properties at Different ages for Each Concrete Strength
(Claeson et al., 2000)
Since the specimens have been cured for just one day, the creep would be larger
than those of the specimens with normal curing process, which are 3 days of steamed
curing or 7 days of moist curing. It is difficult to make this adjustment of the ultimate
creep coefficient values calculated from ACT or CEBFIP. From Figure 4.7, it shows,
Except the at 23 days of loading, the results of analysis with v,, =2.50 match with the data
of test specimen H201 quite well.
64
Figure 4.7 Deflection Results of H201
4.2 Slender Columns Under Sustained Biaxial Bending Loads
4.2.1 Test Specimens
Drysdale et al. (1971) performed an experimental program of sustained biaxial load on
slender concrete columns. There were totally 58 test specimens being tested in this
program. All specimens were the same in column dimension, reinforcing, concrete
strength design, pouring and curing conditions. See Figure 4.8 for the information of test
column dimension and reinforcing.
The following factors were influenced in design of the specimens:
1.Make the scale as large as possible to reduce the size effects.
2.Columns were pinned loaded at both ends with the same eccentricities to provide a
simple and reliable test method.
65
3.
High slenderness ratio was chosen to enlarge creep effects to the total column
deformations.
4.
A square cross section with symmetric reinforcing was chosen to make orientation of
eccentricity selection as less as possible.
Concrete: A standardized mix for a 28 days compressive strength of 4000 psi was
used for all. The mix proportions by weight were: Portland Cement Type 114%, Water
9.1%, Sand46.6%, 3/8" in. crushed limestone30.8%. A slump of 23 in. was obtained.
All tests were loaded at 28 days age of concrete. Instant load tests were completed the
same day while sustained load specimens were subjected under load and along with
control prisms and cylinders in a temperature (75±1°F) and humidity (50±2%) controlled
tent. Standard 6in, diameter by 12in, high cylinders and 5in. by 5in. by10in. prisms
were used to determine the concrete stressstrain relationship.
For shrinkage measurement, 16in, long plain concrete prisms, with 5in. by 5in.
cross sections containing identical reinforcement, were cast with each group. For creep
in concrete, some 32in, long creep prisms with 5in, square cross sections were cast with
some of the early groups of columns. At age of 28 days, the prisms were subjected under
concentric axial load in the same temperature and humiditycontrolled tent. Figure 4.9
shows creep test results of this experiment.
Per both ACI and CEBFIP, the linear creep models are only applied to concrete
stress 0.4f. It is noticed that, when =0.41 at 200 days after loading, the creep
f,
strain 1150x106. Applying Eq. (4.1), one has vu=3.51.
66
Figure 4.8
Column Details of Specimens in Drysdale et al.'s Experiment (Dysdale,
1967)
4.2.2 Test procedure and Equipment
1. Loading Sequences:
The following loading sequences were designed for the experimental tests.
(1)
Short term loading to failure at concrete age of 28 days;
(2)
Constant sustained loading from 28 days of age to failure;
67
(3) For those in (2) not failed for 2 years, unloaded sustained loads and replaced instant
loading until failure.
Figure 4.9 Creep Test Results from Drysdale et al.'s Experiment (Drysdale, 1967)
All tests were performed in identical pairs in order to guaranty the quality of test
data. The columns are classified by Test series according to the end eccentricities, e.
Series A (data ruined) and B, using e=1.00 in at 45 degree from X axis;
Series C, using e=1.00 in at 22.5 degree from X axis;
Series D, using e=1.00 in along X axis;
Series E, using e=1.50 in at 22.5 degree from X axis;
Series F, using e=0.50 in at 22.5 degree from X axis;
See Table 4.2 for the summary of test and analysis results for column test series.
68
2.Strain Measurement:
Demec mechanical gage (8 in. gage length) points were attached on all sides at
five points spaced over the column length. The accuracy of the Demec mechanical strain
indicator is 1x10e5 in/in.
3.Deflection Measurement:
A deflection measurement instrument was used with triangulation to determine
the position at various levels.
4. Test Equipment:
Shortterm load test: 1200 kip capacity, hydraulically operated Baldwin Testing
Machine. Sustained load test: A specially designed loading platform was used.
Figure 4.10a Deflection Results of B2A
4.2.3 Comparison of Test Results with Numerical Analysis
1. Test of Specimen B2A:
69
Figure 4.10b
Strain Diagrams of B2A (strain unit: 1x10
4)
Column B2A was a specimen subjected to biaxial sustained loading for 730
days and failed under an additional instant load. The eccentricity is 1in at 45 degree
from Xaxis. From Table 4.2, the predicted shortterm capacity is 36.2 kips, the constant
sustained load equals 19.218.2 kips.
70
Although ACI20992R model is used to perform the sustained load analysis, it is
still very important to choose a proper ultimate creep coefficient for the computer
analysis program.
Using all original input data for ACI20992R Model, one has:
Per CEBFIP 90 Model: o0=3.62, εcs0=6.47x10e4.
Per creep test: vu=3.51
From Figures 4.10a and b, the analysis using ACI209 Model with vu=3.51 does
not agree with test well when time of loading is over 185 days of loading.
Test of Specimen B3B:
Column B3B was a specimen subjected to biaxial sustained loading for 631 days
and failed under an additional instant load. The eccentricity is 1in at 45 degree from X
axis. From Table 4.2, the predicted shortterm capacity is 36.8 kips, the constant
sustained load equals 23 kips. Thus, ratio of sustained load/load capacity = 0.625.
Per ACI209: vu=1.98, rsh=7.02x10e4;
Per CEBFIP 90 Model: Øo=3.33, εcso=6.47x10e4
Per creep test: 17,
4
=3.51
From Figure 4.11 the results of analysis using ACI20992R Model with v3.51
agrees with those data of test well.
71
Figure 4.11 Strain Diagrams of B3B (strain unit: 1x10 4)
Figures 4.12 and 4.13 display the maximum concrete elastic strain history at mid
column cross section. The test points in the figures are obtained from the recorded
concrete total strains minus the nominal creep strain values recorded from the companion
creep tests. The analysis in both figures do not agree with tests as well as other
comparisons as shown above, since the creep values are more complicated than what they
appear.
Figure 4.12 Concrete Maximum Elastic Strain History of B2A
Figure 4.13 Concrete Maximum Elastic Strain History of B3B
According to Figures 4.12 and 4.13, both tests and analysis results show that the
maximum concrete elastic compressive strain at midcolumn cross sections normally
experiences a drop process due to the stress redistribution to longitudinal reinforcement
73
during the sustained loading period. There is a possibility that the elastic strain becomes
increased (Figure 4.13) while the total deflection keeps increasing.
4.3 Discussions of Selection of Ultimate Creep Coefficient Values
The most important step for an analysis to become successful is to get the right ultimate
creep coefficient value, vu of ACI 20992R Models or the notional creep coefficient
value, Ø0 of CEBFIP Model Code 1990. They are very critical and sensitive to the
theoretical results. In those data verification examples, the test data from the companion
creep tests are used for the analysis. Those values are normally close to the numbers
from CEBFIP Model Code 1990 than ACI 20992R Models. It shows the analysis
results due to using the ultimate Creep Coefficients from the experiments are satisfied.
The reasons are probably due to:
1. The scales of the experimental specimens are relatively smaller than those used in
practical engineering projects. For concrete creep, the smaller the specimen
dimension is, the larger the creep would happen. It seems that the ACI specifications
are established from the specimens tested with relatively larger scale, thus the
ultimate coefficients calculated per ACI129 92R Model are comparatively smaller.
In CEBFIP Model Code 1990, the scales of specimens tested are smaller, thus the
dimension effects are larger. The creep coefficients per CEBFIP are more suitable
for smallscale test specimens. Since the specimens in those slender RC column tests
are in small scale. Thus, CEBFIP Model Code 1990 describes the creep behaviors
for those test specimens better than ACI 20992R Models.
74
2. Concrete creep is a very complicated phenomenon. The current theories are still not
sophisticated enough to describe the comprehensive picture of the phenomenon. In
ACI12992R models, the effects of the concrete strengths have not been included;
the calculation for creep under the high concrete stress is not addressed either, which
are both included in CEBFIP 90. On the other hand, CEBFIP 90 has failed to
include effects of concrete slump and air content that is considered by the ACI129
92R, which might have an impact on the creep development.
CHAPTER 5
DESIGN RECOMMENDATIONS
5.1 Design Examples: Comparison with Current ACI Design Codes
The purpose of this part is to analyze examples and compare with the current ACI
building codes. Under uniaxial bending loads, the effects of slenderness, eccentricities,
reinforcement ratio and concrete strength will be discussed. Under biaxial bending loads,
the effects of the eccentricity orientation and the influence of length over width ratio of
cross section will also be studied. To capture typical engineering situations, the value of
the ultimate creep coefficient 1/, is assumed to equal to 2.0 and the sustained duration is
selected as 30,455 days (83 years). The analysis model is pinnedended and equally
eccentrical at column top and bottom.
5.1.1 Ultimate Strength of Slender Concrete Column under Uniaxial Bending
Loads
Figures 5.1, and 5.2 show the column geometric dimension, material information and P
M strength interaction diagram of a cross section.
1. Effects of slenderness
There are five theoretical specimens included in this group. Keep the same cross
sectional dimensions, longitudinal reinforcement, and concrete strength for all specimens.
At strength boundary of MP diagram, select five different points on the diagram. Let
δb=2 for all specimens.
75
Figure 5.1 Details of Columns in Design Examples
76
Figure 5.2 Short Column PM Interaction Diagram
77
Figure 5.3 shows the results of both computer analysis and ACI318 99 Building
Code. The results show that if columns are relatively shorter, the ultimate strengths
under sustained load calculated per ACI code are closer to those from the present
computer analysis.
2. Effects of eccentricity
Five other theoretical specimens are included in this group. Keep the same
column length, the cross sectional dimensions, longitudinal reinforcement, and concrete
strength for all specimens. Select five different (M, P) strength points in the MP
diagram. Thus, all specimens have different loads and eccentricities at the ends of each
member.
Figure 5.3
Effects of Slenderness
Fig. 5.4 displays the results of analysis from the effects due to eccentricity. It
shows that the larger the eccentrically loaded columns is, the less the buckling strength
78
can be obtained, as per ACI code. This means that the creep influences more to the
column loaded with larger eccentricites.
Figure 5.4 Effects of Eccentricities
3. Effects of reinforcement ratio
Five theoretical specimens are included in this group. Keep the same length of
column, the end eccentricities, the cross sectional dimensions, and concrete strength for
all specimens. Select different longitudinal reinforcement ratio for each member.
Figure 5.5 displays the present computer analysis results. It shows that, the lower
the longitudinal reinforcement ratios in the columns is, the less buckling strength can be
obtained. The results of strengthdecreasing rate vs. reinforcement ratio from present
analysis are larger than the design method as per ACI building code. This could be
explained as: on one hand per ACI building code, when the reinforcement is low enough,
79
the influence of steel would be neglected by using the concrete stiffness only. On another
Enter the password to open this PDF file:
File name:

File size:

Title:

Author:

Subject:

Keywords:

Creation Date:

Modification Date:

Creator:

PDF Producer:

PDF Version:

Page Count:

Preparing document for printing…
0%
Comments 0
Log in to post a comment